Building an Artificial Casting Engineer (ACE) "Building hardware at... the speed of software" remains immensely challenging because the traditional software for engineering, and more specifically manufacturing (Solvers, CAM, etc,) is too damn slow. Introducing Ace-Mini-1, the 1st purpose built AI Casting model from Digital Metal. Ace-Mini-1 is another tiny parameter model fine tuned on casting simulations (filling, solidification, deformation, and more) that cuts an average industry leading solve time of ~65 minutes down to just 0.26 seconds with an average error rate less than 5% across all data. An 8042x speedup in complex CFD with phase change sim time. Ace-Mini-1 allows engineers and agents to close mfg process iteration loops faster, run 100s of simulations in seconds, and optimize geometry and manufacturability of your part.show more

Tesla cut its Gigacasting processing time from 180 seconds... to 75 seconds — nearly 60% faster 🌊 The Model Y Juniper rear casting now weighs approximately 60 kg, down from 67 kg on the previous generation. Much of the industry conversation centers on press size and tonnage. The concrete gains in speed and mass on this high-volume part come from targeted process refinements inside the die and across the production system. -> Processing time reduced from 180s to 75s on the rear casting -> Part weight lowered by 7 kg through incremental design improvements -> Faster cycle achieved while improving microstructure and mechanical properties Conformal cooling makes the difference. Complex water channels, drilled or through 3D-printed inserts directly into the die steel, target hot spots and pull heat out rapidly and evenly across the entire casting. This accelerates solidification, reduces temperature gradients, and allows the part to be ejected sooner without defects. The result is a casting that is both lighter and stronger, produced in less than half the time. Supporting elements include advanced software that controls every injection parameter and runner/gate designs refined through five years of iteration since 2020 + close collaboration among casting designers, die engineers, production, and safety teams running high-volume lines on three continents. Tesla’s manufacturing edge is not the Giga Press hardware itself. It is the accumulated knowledge of how to run these machines at scale. 📊 The Gigacasting Database gives you the full picture of the market: Credit: Atomic Industries - Aaron Slodov ❌ Don't leave your insights to chance with the X algorithm ✅ Subscribe for free to my weekly newsletter about all things Gigacasting and magnesium Thixomolding: 📬show more

🚨 The CrowdStrike 2026 Global Threat Report is here.... In the age of AI, even less sophisticated threat actors can execute complex attacks, and advanced adversaries have become dramatically more dangerous. This year’s report exposes the latest tradecraft of the evasive adversary, who is supercharging attacks with AI and posing an unprecedented threat. Attacks by AI-enabled adversaries increased by 89% in 2025. The average eCrime breakout time plummeted to just 29 minutes. That’s a 65% increase in speed from 2024. And with adversaries using AI to accelerate their attacks and move fluidly across domains, they are evading detection more effectively than ever. Get the latest threat intel findings here:show more

Starting a new project today, building an end to... end system to forecast traffic of flights across cities (starting with Mumbai) The idea is to implement > ingestion service with kafka > data etl with polars > feast for feature store // mlflow for model registry > batch inferencing // dashboard > s3 // postgres for data storage All this orchestrated across multiple DAGs built with Airflow. This year has just been Agents and LLMs all along. Not a bad idea to keep revisiting the traditional format :) Will be posting more of this in the coming days, stay tunedshow more

AN AWS ENGINEER QUIETLY BUILT A 2 PETABYTE HOME... SERVER FOR $9/MONTH THAT KILLS A $3,400/MONTH CLOUD STORAGE BILL the lenovo thinkstation pgx ships nvidia's gb10 grace blackwell superchip and 128gb of unified memory in a box the size of a mac mini at 1.2kg it runs an 80b qwen3 coder model at 25 to 40 tokens per second and a 196b step-3.5-flash moe model at 20 tokens per second locally the gb10 packs 6,144 cuda cores, 192 fifth-generation tensor cores and rates at 1 petaflop of fp4 with sparsity from a single 240 watt usb-c power supply fine tuning qwen 2.5 7b with lora took 18 minutes and 41gb of unified memory while the gpu pulled 65 watts and peaked at 77 degrees the box pulls a docker container from nvidia's registry and serves a frontier model on your local network with tool calling and zero data leaving your desk bookmark this and read the article belowshow more

Fine-tune DeepSeek-OCR on your own language! (100% local) DeepSeek-OCR... is a 3B-parameter vision model that achieves 97% precision while using 10× fewer vision tokens than text-based LLMs. It handles tables, papers, and handwriting without killing your GPU or budget. Why it matters: Most vision models treat documents as massive sequences of tokens, making long-context processing expensive and slow. DeepSeek-OCR uses context optical compression to convert 2D layouts into vision tokens, enabling efficient processing of complex documents. The best part? You can easily fine-tune it for your specific use case on a single GPU. I used Unsloth to run this experiment on Persian text and saw an 88.26% improvement in character error rate. ↳ Base model: 149% character error rate (CER) ↳ Fine-tuned model: 60% CER (57% more accurate) ↳ Training time: 60 steps on a single GPU Persian was just the test case. You can swap in your own dataset for any language, document type, or specific domain you're working with. I've shared the complete guide in the next tweet - all the code, notebooks, and environment setup ready to run with a single click. Everything is 100% open-source!show more

Today marks General Availability of AgentCore, a set of... infrastructure building blocks for developers and companies to build secure, scalable agents. When we first started AWS, the vast majority of developers were spending most of their time on the undifferentiated heavy lifting of infrastructure instead of what differentiated their feature. So, we solved that problem by building primitive building blocks like compute and storage and database that would allow teammates and customers to quickly build and deploy new experiences without having to reinvent the wheel each time. We realized the same thing was happening with AI agents. It's too difficult and it's slowing customers down. That's why we created AgentCore, a set of services to build, deploy, and operate highly capable agents using any framework or model, with enterprise-grade security and scalability. These building blocks (like serverless secure runtime, memory, observability, a gateway that does MCP translation, etc) help customers tackle some of the biggest challenges of going from prototype to production, much more quickly, securely, and scalably. AgentCore has been in preview for several weeks, and customers have been quite excited about it. The AgentCore SDK has already been downloaded over a million times and we're seeing transformative results, such as Cohere Health expecting to reduce medical review times by 30-40% in highly regulated healthcare, and teams at Cox Automotive and Experian are embracing its flexibility to deploy and operate agents at scale. Inside Amazon, our Amazon Devices Operations & Supply Chain team is using AgentCore to develop an agentic manufacturing approach where AI agents work together to automate manual processes – turning what used to be days of engineering time into processes that take under an hour with high precision. Just like AWS changed how companies build and scale applications, we believe AgentCore will do the same for AI agents, enabling the next generation of innovation.show more

🚨 AMERICA JUST BUILT THE WORLD’S MOST POWERFUL METAL... 3D PRINTER AND IT’S ABOUT TO MASS-PRODUCE ROCKETS AND MISSILES. Divergent Technologies has unveiled the Monolith One, a giant industrial metal printer standing over 8 meters tall and armed with 12 high-powered lasers delivering a combined 24 kilowatts of energy. Unlike typical 3D printers used for prototypes, this machine is built for serious, high-volume production. It can print large, complex aerospace and defense parts in aluminum, titanium, steel, and nickel alloys and it roughly doubles the output of current systems. Why this matters: • Divergent plans to install 64 more of these machines in a massive new 430,000 sq ft factory in Long Beach, California • Once running, the facility aims to produce tens of thousands of munition airframes per year plus hundreds of thousands of critical metal components • It slashes manufacturing time from months down to weeks or even days • The company already supplies major players like Lockheed Martin and RTX The deeper implication: This isn’t just another 3D printer. It represents a shift toward software-defined, on-demand manufacturing at industrial scale for mission-critical hardware. As defense and aerospace demand skyrockets, traditional supply chains are too slow. Systems like Monolith One could become a cornerstone of faster, more resilient domestic production especially for complex structures that are difficult or impossible to make conventionally. We’re watching the industrialization of additive manufacturing in real time. How do you think large-scale 3D printing will change aerospace and defense manufacturing over the next decade? Follow for more frontier manufacturing and defense technology.show more

Introducing Poetic: a new AI system that executes complex... multi-hour tasks with 99%+ accuracy and 10x fewer tokens than agents. We raised $50M at $500M from Kleiner Perkins, Founders Fund, First Harmonic, and Genius Ventures to build AI that does complex work inside Fortune 500 companies without hallucination. While code is too brittle, agents are too unpredictable. The work that runs the global economy - anti-money laundering, fraud investigations, underwriting - needs extreme accuracy. So we built a new kind of software that pairs the flexibility of AI with the predictability of code. When the world stays the same, Poetic runs fixed code: fast, cheap, identical every time. When the world changes, Poetic uses AI to regenerate its approach and find its way back to the objective. In one year, we went from zero to an eight-figure run rate as a team of four. Since then, we’ve scaled the team and executed the highest-stakes processes at AIG, SoFi, and Chime. At SoFi, a large US bank, Poetic reached 99%+ quality on fraud investigations in five weeks.show more

Suppose two cars are traveling on an expressway at... 100 km/h with a 50-meter gap. If the front car slams on its brakes, it takes about 3.5 seconds and nearly 50 meters to stop. If the driver behind does nothing, a collision occurs in exactly 3.5 seconds. This gives the rear driver a 3.5-second window to avoid a crash. For an alert driver, 1 second is reaction time, leaving 2.5 seconds to brake and stop safely. Now, consider the same 50-meter gap, but the front car is reversing at 10 km/h. For the car approaching from behind at 100 km/h, the closing speed skyrockets to 110 km/h, slashing the time before collision to just 1.6 seconds. In reality, it is even worse: because drivers never expect a vehicle to reverse on an expressway, their brains freeze. Reaction time jumps from 1 second to 2, 3, or more. At high speeds, drivers often crash before even realizing they need to brake. That's why reversing on an expressway is hundred times more dangerous than a sudden braking. A maneuver that feels slow to the reversing driver creates an unexpectedly violent closing speed for those behind, leaving virtually no time to react. In the video, a family of seven missed their exit on the Dehradun expressway to Haridwar. They reversed, were struck from behind, and four of them died.show more

Everyone’s now empowered to ship software swiftly. That's no... longer the hard part! The real difficulty comes with building something that still matters once the "I built this with AI" novelty is done. Replit CEO and Founder Amjad Masad has been living and breathing that problem, and he’s bringing that perspective to the Disrupt Stage at Disrupt 2026. It's one of more than 200 sessions this year built around a single pivotal question: How do you build an enduring company in the AI era, not just a fast one? The current ticket pricing ends tonight at 11:59 p.m. PT. If Disrupt's been on your radar, now is the time to register before rates increase:show more

Zuckerberg built his own AI agent to run Meta.... this man is literally becoming Tony Stark. it pulls data from every team inside the company so he can skip meetings, skip the chain of command, and make decisions faster than any human process allows. 78,000 employees have their own AI agents now too. one messages coworkers on your behalf. another acts as your AI chief of staff. their agents talk to each other in an internal network. humans optional. Meta also bought an entire social media platform built for AI agents to interact with each other. read that again. Zuck said he wants every person at Meta to have a personal AI agent. then every person outside Meta. the Jarvis era started.show more

this is f*cking gold engineers at Meta just deleted... the most expensive part of multi-agent systems: instead of training a communication topology, they compile a fresh one for every query 20 agents went from 7 hours to 6 minutes. the problem everyone hits: 5 agents works. 20 agents turns into a group chat that answers slower than one model and costs more than the task is worth. ReActNet's fix is that the graph is written per query, not learned once: > an LLM controller reads the query and the agent roster > it compiles a sequence of directed graphs, one per reasoning stage > every edge carries a written instruction, e.g. "list boundary cases for this behavior" > each agent updates its state from its own previous state plus assigned neighbors > a final node aggregates all five states into the answer > no RL, no gradients, no training stage at all what that buys on gpt-4o: 92.75 average across 6 benchmarks, best on 5 of them. 100.00 on MultiArith. 92.74 pass@1 on HumanEval, +21 over a single model. and the number that should worry anyone running a swarm: at 20 agents GPTSwarm needs 412 minutes and $41.42. ReActNet needs 6.22 minutes and $6.53 and scores higher. the honest catch: more agents did not make it smarter. 5 agents scored 79.74, 20 scored 77.77. the topology was never the thing to learn.show more

i live on ranch, and it cost me less... than buying an apartment in california. the downside is isolation from city life. if that doesn't bother you, quality of life shoots up. you'll become obsessed with animals 😂 i have 2 mini cows, mini horses, poodles. must stop, but... learning to train them is too fun. i've included a video of my mini cow giving "hello friend" head nods to my new mini horse i'm also building a soundshed on the property with two hifi audio systems to incept my friends to fall in love with the hobby. over time, you also become interested in being off grid. think solar, starlink, rain catchment, cattle + chickens, herb + fruit + veggie garden. life slows down and you start to value building these skills—to self-sustain. going outside between Zoom calls and just looking at the trees and hearing wind rustle through them is so calming. this is how i grow up. shouldn't have done a full 15 years in the city in between. i've built a bunch of guest cabins, and friends can come stay for a couple weeks at a time. i put guitars + pianos + hifi in their cabins so they can feel creative next up is a giant golf cart racing course that traverses through the propertyshow more

Introducing RL Environment Creator Skill Now any one can... create RL environments $ npx skills add adithya-s-k/RL_Envs_101 > You can create environments across multiple frameworks like OpenEnv, OpenReward, Verifiers, NemoGym ... > the repo has live working examples of environments that your coding agent can reference > The skill is design to first understand what type of model you are training and create an environment while keeping that in mind ps. There’s a lot more to building RL environments that can be used for training. One major aspect is the data, which this skill can’t directly solve. However, the skill will help with implementing tools, rewards, and other components of an RL environment, making it easier to go from idea to implementation quickly across different frameworks. Let me know if you’d be interested in a detailed, end-to-end blog/tutorial on building an environment and actually training a model for a useful use case.show more

🧃 Introducing stereOS: a Linux based operating system hardened... and purpose built for AI agents. It's clear that agents need an ACTUAL operating system (not what people are calling an "OS") to witness the full breadth and depth of their capabilities while mitigating the blast radius of autonomous, untrusted actors. But there are so many problems with AI sandboxes today: * Going out to the apple store and buying a mac mini will never scale and is way too expensive (obviously) * Running in Docker is too restrictive (agents can't stand up their own container infrastructure, no sub virtualization, docker-in-docker is very broken) * Firecracker strips all the hardware so GPU PCIe passthrough, secure boot, FIPs, etc. is out of the question. * Native VMs are too fat and the overhead of 1 agent per VM is too much. stereOS takes a different approach: it's a full NixOS system that you boot and then kick off agent sandboxes inside with gVisor + /nix/store namespace mounting. Each agent gets their own kernel and the /nix/store is read only by nature. Even if the agent was somehow able to escape the gVisor virtual kernel, they'd land on the NixOS system as the "agent" user! Not your actual hardware!! If you want to take a defense-in-depth approach, we support "native" agents that run at the system level kicked off by our `agentd` utility. These agents, on their own, can manage and kick off other sub agents using the internal sandboxing mechanisms. Today, we're open sourcing all of this: * stereOS: our purpose built Linux OS - * masterblaster: client utility to launch, manage, and orchestrate agents - * stereosd: the stereOS system control plane daemon - * agentd: the stereOS system agent management daemon - Give it a try, throw us a star, and let me know what you think 🧃⭐️show more

A machine like this can cost $500,000 to well... over $1 million to make parts that may be worth only $10-15 This is the real manufacturing story. This INDEX six-spindle automatic carries 6 motorised spindles, up to 12 CNC tool carriers, operates at 8,000+ rpm, and weighs 7.2 tonnes. Different operations happen simultaneously as the spindle drum indexes each workpiece from station to station. A real scale production example produced a precision component in 11 seconds, versus 38 seconds on a conventional single-spindle lathe. That's roughly 327 parts per hour before downtime. But the machine is only the hardware. Tool geometry, CNC programs, cutting parameters, spindle synchronisation, tooling, bar feeding, chip evacuation, coolant, inspection and collision checked simulation all have to be engineered around the exact part. That is what it takes to integrate this machine into a factory workflow. This accumulated capability is what kept Germany and Japan at the pinnacle of machine-tool manufacturing for decades almost unchallenged. The advantage wasn't just building the hardware, but knowing how to program, tool and integrate these machines for thousands of different manufacturing requirements globally. China has now built much of that ecosystem at extraordinary scale, machines, controls, tooling, software, automation and integration. Its huge domestic manufacturing base has accelerated that learning curve dramatically. Today, Chinese manufacturers can increasingly offer sophisticated CNC and multi-spindle systems at 30-40% lower total costs in most applications, putting serious and relentless price pressure on German and Japanese builders. China produced 37% of the world's machine tools in 2025, compared with 12% for Germany and 10% for Japan. These are the machines that make the machines and ultimately determine how much an economy can manufacture. China is the biggest player as of now and growing faster than anyone else in manufacturing high-end machining tools. Source, Daniel Janssonshow more

HTML Artifacts are a big part of how I... work with agents now. Artifacts can be more than just static files. When combined with agents, they can take action or help you take action. This unlocks all kinds of interesting ways to work with agents. This is clearly the future. Check out this writing and scheduler artifact I built in a few minutes. It uses a bit of HTML and JS. All the data is in markdown (Obsidian vaults), so the agent can access and modify it at any time. No DB needed. No sophisticated functionalities. The agent decides all that for me based on the skills, context, and memory it has access to. The best part about this simple stack is that all the important information stays with me. This has allowed me to build a recursive self-improving system and automations that can better tap into coding agents like Codex or Claude Code. I could have paid or built an entire app for scheduling posts, and there are so many of them out there. But I don't need to. I've realized a simple artifact does the job. And the simplicity of it is actually an advantage. Very little maintenance for very high returns on personalization, time, and efficiency. The other benefit of this is that I can add features as I please. That level of personalization feels magical, and we should all be pursuing more of it. All of this just keeps compounding. Of course, this example is just about writing. But I have similar artifacts for research, design, experimentation, evaluation, and so much more. And no, I didn't actually publish the post example I shared in the clip. It was just for demonstration purposes. I actually spend more time than this when writing together with agents. Lastly, having built my own agent orchestrator tool has made me realize that simplifying the tool stack is a superpower. If you are curious about how all this works, I will do a live session next week:show more

Elon Musk gave the entire entertainment industry its expiration... date, and he is the one building the thing that kills it. Musk: “My guess is that we see the first compelling half hour, pure AI show next year.” Next year. A complete show generated entirely by AI. No writers. No actors. No cameras. No sets. No crew. No studio. Just a prompt and enough compute to render a reality that never physically existed. And shows are the easy part. Musk: “I say probably we’re maybe three years away from AI does the whole video game.” A show plays the same way every time. A game has to generate a living world that reacts to every decision in real time across every single frame. That is a fundamentally harder class of problem. And Musk put three years on it. Right now a single AAA title takes seven years and half a billion dollars across thousands of engineers and artists just to ship it. Musk is describing a world where one person types a paragraph and gets something comparable. The entire value proposition of a multi-billion dollar industry lives inside that gap. And it closes in thirty-six months. But the prediction is not the story. The person making it is. This is not an analyst speculating from the sidelines. This is the man building the largest AI compute clusters on the planet. The man who built xAI from zero in under two years. The man stacking hundreds of thousands of GPUs into facilities designed to do exactly what he is describing. When Musk says three years, he is not guessing about what someone else might eventually ship. He is reading you a delivery date off his own roadmap. Every media company on Earth is valued on a single assumption. That quality content is expensive and difficult to produce at scale. That one assumption is the structural foundation underneath every studio, every network, and every publisher in existence. Musk is dismantling it with raw compute. The studios still parading thousand-person production teams are not demonstrating strength. They are advertising the exact cost structure that one person with a prompt and a GPU allocation is about to make irrelevant. And it does not stop at entertainment. If AI can generate an interactive world that responds to human input in real time, it can generate anything. Advertising. Architecture. Training simulations. Product design. Every industry built on humans manually constructing visual experiences frame by frame is sitting on the same countdown Musk just read out loud. Now zoom out. Because this is not just an industry story. For the entire history of human civilization, the distance between imagining a world and actually creating one required thousands of people, millions of hours, and billions of dollars. That distance built Hollywood. That distance built the gaming industry. That distance made content scarce and studios powerful. Musk is collapsing that distance to zero. When the gap between imagining something and it existing disappears, every business model built on the difficulty of creation disappears with it. That is not disruption. That is a full inversion of how human beings create. Musk did not make a casual prediction on that podcast. He told you what he is building. He told you the timeline. And he told you which industries do not survive it. The entertainment industry is still debating whether this future is real. Musk is not part of that debate. He is building. And he just told you the delivery date.show more

eXoZymes (eXoZymes) (Nasdaq: EXOZ) CEO Michael Heltzen was featured... on Session 20 of the GeneCoda podcast Executive Insights for Life Sciences Innovators, in an episode titled “AI-enabled cell-free biomanufacturing and the future of enzyme engineering,” streamed live on YouTube on April 17. eXoZymes is a B2i Digital (B2i Digital) Featured Company. See their full profile at In the conversation, Heltzen discusses the company’s thesis that producing complex molecules outside living cells can enable greater control, scalability, and speed compared to traditional biologic manufacturing. The discussion also covers how AI is helping identify and optimize enzyme pathways, the strategic considerations involved in building a platform company in today’s capital environment, what differentiates cell-free systems from established synthetic biology approaches, and where cell-free biomanufacturing could create commercial impact across nutraceutical, pharmaceutical, and industrial markets. Watch or listen here: Separately, eXoZymes and Cayman Chemical were featured in a Springwise case study on how the two companies are rethinking chemical manufacturing. The piece focuses on eXoZymes’ work bringing enzyme pathways from biology into the world of chemistry, with Cayman Chemical helping to scale the approach. CCO Damien Perriman is featured in the article. See the eXoZymes summary: eXoZymes is a Los Angeles, California-based biotechnology company that has developed a biomanufacturing platform offering the tools and insights to design, engineer, control, and optimize nature’s own natural processes to produce highly valuable natural products via a commercially scalable, sustainable, and abundant alternative: exozymes. eXoZymes is led by an experienced management team including Michael Heltzen, CEO; Damien Perriman, CCO; Dr. Tyler Korman, CSO and Co-founder; Dr. Paul Opgenorth, VP of Development and Co-founder; Fouad Nawaz, VP of Finance; Lasse H. Görlitz, VP of Communications; and Amy Lunzer, Chief of Staff. Learn more about eXoZymes Inc. at For investor inquiries, visit and learn about other B2i Digital Featured Companies at Disclosure: David Shapiro, Chief Executive Officer of B2i Digital, personally purchased in the open market and currently owns shares of unrestricted EXOZ stock, in line with B2i Digital’s practice of investing alongside its Featured Companies. This post is not intended to solicit the sale of EXOZ or any security, and it is not intended to offer any opinion on EXOZ as an investment. Conduct your own research and consult with your own professional advisors prior to making any investment decisions. See the complete disclosure in the Risks and Disclosures section ofshow more

😓 Air India 🇮🇳 Flight AI171 with fully loaded... Boeing 787-7 Dreamliner fatal accident: I‘m an airline pilot with >15‘000h of experience and a physics institute: My brief PRELIMINARY analysis of the visible facts from the video of the takeoff: * The flaps are only slightly extended, presumably to position 1 instead of 5. * The landing gear is still extended, which should have been retracted at this altitude and causes additional drag. * The aircraft is at a high angle of attack, which confirms the insufficient flap setting. * From the video and witness accounts, only low engine noise is audible. * Neither smoke nor fire is visible. * An engine failure is less likely. The most probable cause is presumably a human factor, an incorrectly chosen, insufficient flap setting for takeoff, and consequently an inadequately selected thrust. In this context, the correlated speeds were too low because they were calculated for a larger flap setting or a lighter aircraft. As a result, the aircraft took off with insufficient speed and intentionally but falsely derated thrust, was therefore on the unstable side, and rapidly lost more speed and altitude due to the additional failure to retract the landing gear in a timely manner, leading to a subsequent stall at low altitude and crash. For the experts: the aircraft got onto the wrong side of the speed vs drag curve and maneuvered itself into a corner from where there is no escape. Another possible cause could also have been an incorrect input of a wrong takeoff weight into the Flight Management System, resulting in too low thrust and too low speeds. The pilots got startled after takeoff, couldn’t wrap their head around what went wrong and incorrectly prioritized making an emergency call instead of flying the aircraft first, manually increasing thrust immediately to maximum, and retracting the landing gear. In summary of this very early and preliminary assessment (your confidence level should be as low as mine): The most probable cause is human error 😓 - as most of the time these days. Not because the pilots got worse (although that effect can be observed as well with prioritization of diversity over competence) - but because technology got so much better.show more